MFIX Home Monday, May 12
Home | Contact us
MFIX Home > Announcements > MFIX Announcements
NETL Home
MFIX Home
About MFIX
Announcements
Partners
Documentation
Publications
Results
Links
Download and Members Area
Open Citations
DOE Home
ORNL Home

Multiphase Flow with Interphase eXchanges

Current Announcements:

April 27, 2008 "Paper on open-source experience with MFIX"

April 27, 2008 "Paper based on open-source experience with MFIX;


Dear MFIX user,

The following paper discussing the open source experiences with MFIX has been published: Madhava Syamlal, Thomas J. O'Brien, Sofiane Benyahia, Aytekin Gel, and Sreekanth Pannala, Open-Source Software in Computational Research: A Case Study, Modelling and Simulation in Engineering, vol. 2008, Article ID 937542, 10 pages, 2008. doi:10.1155/2008/937542. You may access this article from the Table of Contents of Volume 2008, which is located at the following link: http://www.hindawi.com/journals/mse/volume-2008/ Alternatively, you may directly access your article at the following location: http://www.hindawi.com/getarticle.aspx?doi=10.1155/2008/937542

Back to Top

December 31, 2007 "MFIX2007-4 Release Announcement"

December 31, 2007 "MFIX2007-4 Release Announcement;


Dear MFIX user,

We are pleased to announce the release of the new stable version of MFIX (MFIX2007-4). Please visit http://www.mfix.org/members/download.php to download the latest version. Please see the attached release notes that highlight the changes from the previous version.

Please send any software related questions to mfix-help@lists.mfix.org or administrative questions to admin@lists.mfix.org.

Regards,

MFIX Development Team

Changes from MFIX2006-4 to MFIX2007-4 (Date: 12/31/2007)

Changes in existing features:

  • Significant bug fixes were made to the DEM model implemented in MFIX (see log listing at [1]). Most of the modifications were made by Rahul Garg at Iowa State during his summer internship at NETL. Here is a brief summary of the changes:
    • Number of bug fixes and enhancements were made to handle general multiparticle mixtures with varying density as well as particle size
    • The code has tested 2D and 3D capability
  • Added/modified some test cases and tutorial cases. The test and tutorial cases are benchmarked and the summary tables are located in respective directories.
  • Several bug fixes and please have a look at the log history for a more detailed list of changes at [2]. Some notable ones are:
    • Fixed an error in the calculation of the enthalpy transfer because of mass transfer
    • The divergence terms are called consistently when solids viscosity is a constant
    • Bug fixed for wall cells that may be tagged as fluid in cyclic cases
    • Changed old-time velocity V_s0 to current velocity U_s, V_s and W_s in calc_s_ddot_s. This will accelerate convergence.
    • Moved read_database call from check_data_01 to check_data_06
  • Updated equations document [3]
  • Updated Readme - part of the distribution and readily available from [4]
New features added:
  • The following additions were made to the DEM code [5]:
    • A new and efficient cell linked list based search algorithm has been implemented and tested in 3-D DEM simulations
    • Added an interpolation suite that can be turned on to calculate the individual drag forces on the particles based on particle locations rather than cell averages
    • Added a small tool to generate input deck for DEM cases for multiparticle mixtures in mfix/tools/DES_Particle_Generator/
  • Make file has following additions
    • Option to compile using g95 under Cygwin
    • Modified the selection table under 32 bit compilers, instead of multiple questions for versions of Intel compiler etc., everything is under single list similar to 64 bit compiler options
    • Have an option to link dgtsv if machine optimized libraries are available
    • The make_file has additional options for CRAY XT3/XT4 systems
  • Modifications to make an optimized scalable version of MFIX for large parallel machines
    • The residual calculations are collected into 4 calls for every iteration
    • The dot-products in BiCGStab are reduced by almost 50%
    • Additional preconditioning (alternating sweep directions) and solver options (CG, SOR) have been added
    • Additional tools to collect statistics about scalability and load-balance
References
  1. http://www.mfix.org/cgi-bin/cvsweb.cgi/mfix/model/des/#dirlist
  2. http://www.mfix.org/cgi-bin/cvsweb.cgi/mfix/model/#dirlist
  3. S. Benyahia, M. Syamlal, T.J. O’Brien, “Summary of MFIX Equations 2005-4”, From URL http://www.mfix.org/documentation/MfixEquations2005-4-3.pdf, July 2007.
  4. http://www.mfix.org/documentation/Readme.pdf
  5. The MFIX-DEM document is currently being updated and the DEM mailing list will be informed as soon as the document is available.

    6. Back to Top

    July 19, 2007 "MFIX won the 2007 R&D 100 Award"

    July 19, 2007 "MFIX won the 2007 R&D 100 Award;


    Since 1963, R&D Magazine has annually awarded R&D 100 Awards to the 100 most technologically significant new products to hit the market during the year. According to R&D Magazine, the goal of the award is to spotlight major breakthroughs in products and processes with the capacity to improve the standard of living for many people. Read more about it at http://www.netl.doe.gov/newsroom/features/07-2007.html#mfix or http://www.fossil.energy.gov/news/techlines/2007/07058-FE_Technologies_Win_Awards.html

    Back to Top

    December 14, 2006 "MFIX2006-4 Release Announcement"

    December 14, 2006 "MFIX2006-4 Release Announcement"


    Dear MFIX user,

    We are pleased to announce the release of the new stable version of MFIX (MFIX2006-4). Please visit http://www.mfix.org/members/download.php to download the latest version. Please see the attached release notes that highlight the changes from the previous version.

    Please send any software related questions to mfix-help@lists.mfix.org or administrative questions to admin@lists.mfix.org.

    Regards,

    MFIX Development Team

    Changes from MFIX2005-4 to MFIX2006-4 (Date: 12/14/2006)

    Changes in existing features:

    • Significant changes were made to the DEM model implemented in MFIX and the current implementation is documented in a detailed report [1]. The changes not only improve the efficiency of the code but also accuracy to some extent. Here is a brief summary of the changes:

      • The neighbor search algorithms have been improved so that one can perform neighbor search after a prescribed number of times or when the particles move out of a prescribed halo region of influence.

    • The quadtree/octree is not initialized at every iteration but only at user-specified interval and when the maximum number of quads is reached.

    • Additional optimizations to the quadtree were made which are described in the DEM document [1]

    • Particle search within fluid cells has been drastically improved by taking advantage of the Cartesian topology and looking only at the neighboring cells based on the local particle velocity.

    • All the information being passed from DEM to MFIX in coupled simulations was not consistently adhering to the staggered grid formulation and this has been fixed. In addition, there were other bugs related to the coupling and are fixed. For more details, probe the CVS log at [2]

    • All geometric inputs and other duplicate initializations for the discrete part are removed and the information would be obtained from MFIX inputs.

    • All the arrays have been made allocatable so that better management of the memory can be made.

    • The array indexing has been swapped so that the longest index is in the first location so that better cache utilization can been accomplished during calculations.

    • The code has been extensively cleaned and several alternate computations were constructed to improve the overall efficiency of the code.

    • A major bug related to transferring void-fraction information from DEM to MFIX continuum side has been fixed.

    • The code was additional commented, formatted and made more modular for easy readability.

    • Many variables are renamed for easy understanding and consistency.

    • The code has tested 2D and 3D capability.

    • The interphase coupling inconsistency reported by Jin Sun at Iowa State is fixed and details are available in DEM document's appendix [1]

    • Added/modified some test cases and tutorial cases. The test and tutorial cases are benchmarked and the summary tables are located in respective directories.

    • Minor Open-MP bug fixes

    • A major bug related to NFS writes which was causing premature exits for parallel jobs has been fixed

    • One can look at the log history for a more detailed list of changes at [3]

    New features added:

    • The following additions were made to the DEM code and refer to the report [1] for more details:

      • The DEM output is written out in VTK format at the same frequency as the void fraction for coupled simulations and at user-specified frequency for DEM only simulations. One could now superpose DEM particle visualization on top of continuum field information in Paraview.

    • Added restart capability to DEM.

    • Added a small tool to generate input deck for DEM cases in mfix/tools/DES_Particle_Generator/

    • The adjustable time sequence is restored so that the simulation time depends on the characteristic time-scales rather than something imposed artificially. This seems to accelerate the simulations considerably.

    • Newton’s 3rd Law is implemented for force calculations to eliminate the duplicate force calculation for any given pair.

    • Iddir & Arastoopour's [4] kinetic theory for multiparticle types was added.  Available via MFIX input KT_TYPE. Refer to documentation [5] for more details.

      • Added a tutorial case for this model

    • Added several new radial distribution functions. Available via a new MFIX input RDF_TYPE.

    • Added a small tool to the mfix/tools/MFIX_Grid_Generator/ to generate grid input to MFIX for more complicated cases with non-uniform grid.

    • Added functionality to do both hyperbolic tangent as well has sigmoidal functional blending of the Schaeffer/Kinetic theory stresses.

    References

    1. http://www.mfix.org/documents/MFIXDEM2006-4-1.pdf

    2. http://www.mfix.org/cgi-bin/cvsweb.cgi/mfix/model/des/#dirlist

    3. http://www.mfix.org/cgi-bin/cvsweb.cgi/mfix/model/#dirlist

    4. Iddir, Y.H. and H. Arastoopour, Modeling of multitype particle flow using the kinetic theory approach. AIChE Journal, 2005. 51(6): p. 1620-1632.

    5. http://www.mfix.org/documents/MFIXIAEqns2006-4-1.pdf



    Back to Top

    "October 10, 2006 "MFIX team won FLC technology transfer award"

    MFIX team members won the Federal Laboratory Consortium mid-Atlantic regional award for excellence in technology transfer for 2006 (see the announcement). The award was presented to Dr. Madhava Syamlal of National Energy Technology Laboratory on September 21, 2006. The winning team also included Dr. Thomas J. O'Brien (DOE), Dr. Chris Guenther (DOE) and Dr. Sofiane Benyahia (Fluent Inc.) of National Energy Technology Laboratory and Dr. Sreekanth Pannala of Oak Ridge National Laboratory.

    The Federal Laboratory Consortium (FLC) for Technology Transfer is the nationwide network of federal laboratories that provides the forum to develop strategies and opportunities for linking the laboratory mission technologies and expertise with the marketplace. The Mid-Atlantic Region of the FLC comprises over 70 Federal laboratories and agencies. The awards are given to employees of FLC MAR member laboratories for their effort in transferring technologies to the marketplace.

    Back to Top

    "July 25, 2006 "The VisIT (v 1.5.3) visualization software now supports MFIX data format"

    The VisIT (v 1.5.3) visualization software now supports MFIX data format. See http://www.llnl.gov/visit/new.html . A quick guide to get you started is at http://www.mfix.org/members/develop/visit/VisitMFIXTutorial.pdf.

    Back to Top

    "April 3, 2006 "A paper discussing the open source experiences with MFIX has been posted"

    A paper discussing the open source experiences with MFIX has been posted in open citations. The paper will be presented at the Fifth World Congress on Particle Technology, Orlando, Florida, April 23-27, 2006. It discusses the results of two user surveys and the experiences of MFIX team.

    Back to Top

    January 26, 2006 "Summary of the current set of equations in MFIX"

    A document summarizing the current set of equations in MFIX has been posted at http://www.mfix.org/documentation/MfixEquations2005-4.pdf. The document will be updated when the equations in MFIX are revised or errors in the document needs to be fixed. Every posted version of the document will be maintained with a fixed URL so that these documents may be cited in publications.

    Back to Top

    December 13, 2005 "MFIX2005-4 Release Announcement"

    December 13, 2005 "MFIX2005-4 Release Announcement"
    Dear MFIX user,

    We are pleased to announce the release of the new stable version of MFIX (MFIX2005-4). Please visit http://www.mfix.org/members/download.php to download the latest version. Please see the attached release notes that highlight the changes from the previous version.

    Please send any questions to mfix-help@lists.mfix.org or admin@lists.mfix.org.

    Regards,

    MFIX Development Team

    Changes from MFIX2005-3 to MFIX2005-4 (Date: 12/12/05)

    Changes in existing features:

    • An error in the energy equation introduced in the last version was corrected.
    • Added/modified some test cases and tutorial cases.
    • Several more updates for code clean up and for bug fixes - please refer to the CVS [1]

    New features

    • A recently published thermochemical database of Burcat and Ruscic (2005) was linked with MFIX. This gives MFIX users access to accurate thermochemical data for over 1300 species, which includes specific heat, enthalpy, entropy, free energy, formation enthalpy and molecular weight. In addition to the improvement in the accuracy of the calculations, the linkage with the database eliminates the potential for errors when data are entered manually. The database authors’ permission was obtained for including it in MFIX; the database can be used only for noncommercial work (see the notice in the database) [2]. The rate of reaction term is also computed automatically now. See the readme file for details.
    • A reader for the Paraview visualization software and instructions for using the software has been added.

    References

    1. http://www.mfix.org/cgi-bin/cvsweb.cgi/mfix/model/#dirlist
    2. A. Burcat and B. Ruscic, "Third Millennium Ideal Gas and Condensed Phase Thermochemical Database for Combustion with updates from Active Thermochemical Tables" ANL-05/20 and TAE 960 Technion-IIT, Aerospace Engineering, and Argonne National Laboratory, Chemistry Division, September 2005.)

    Back to Top

    December 13, 2005 "MFIX2005-3 Release Announcement"

    Dear MFIX user,

    We are pleased to announce the release of the new stable version of MFIX (MFIX2005-3). Please visit http://www.mfix.org/members/download.php to download the latest version. Please see the attached release notes which highlight the changes from the previous version.

    This e-mail also serves as a reminder that everyone who has become a member to download the code cannot re-distribute the code or pass the username and passwords to other people.

    Please send any questions to mfix-help@lists.mfix.org or admin@lists.mfix.org.

    Regards,

    MFIX Development Team


    Changes from MFIX2004-3 to MFIX2005-3 (Date: 09/12/05)

    Changes in existing features:

    • The iteration scheme was modified substantially to improve the accuracy of the calculations.  See [1] for implementation details and [2] for verification of momentum conservation.  If you have a customized version of iterate.f you will need to transfer your changes into the new iterate.f. No other user action is needed. 
    • A new method for normalizing the residuals of pressure and solids volume fraction correction equations was added.  See [3] for details and [4] for verification of technique.  The new normalization method can be activated by setting Norm_g = 0.0 and Norm_s = 0.0 in mfix.dat.  The use of this feature is recommended because it speeds up calculations, especially near a steady state or even in transient simulations.
    • The keywords C(2) and C(3) used for calibrating  Syamlal-O’Brien drag correlation have been replaced with drag_c1 and drag_d1.  (The C-array is reserved for user-defined parameters and should not be used as permanent keywords.)
    • The drag coefficients, interphase heat transfer coefficients, specific heats, conductivities and diffusivities are now calculated every iteration (implicit method), unlike the previous method of calculating them every time step (explicit method).  The calculation of the gas-solids drag coefficient may be underrelaxed with the key word UR_F_gs.  The default value of UR_F_gs is 1; setting it to 0 will give the old method for gas-solids drag calculation.  To use explicit method for quantities set the appropriate flags to .FALSE. in iterate.f
    • Added the capability to specify different preconditioners (LEQ_PC) for different sets of equations.
    • The time-step size will be reduced, if the maximum number of iterations is exceeded.
    • Added/modified some test cases and tutorial cases.
    • Makefile modified to support Intel's Ifort 9.0 compiler and to pick up user-modified files in the subdirectories under model directory. For example, Fortran and include files in run_directory/dqmom would be used to override the corresponding files in mfix/model/dqmom directory during compilation. Support for AMD Opteron systems is also added.
    • Scripts and changes to aid easy compilation with windows visual Fortran compiler.
    • Updated benchmark files (tests/MFIX_Tests_Benchmarking.html and tutorials/MFIX_Tutorials_Benchmarking.html) to reflect the current test/tutorial cases along with reference wall clock times.
    • Make all PEs to write error message to a log file when a parallel run is aborted. This is to ensure logs are written for fatal errors even when dmp_log is set to false.
    • New animate mfix (Version 1.3) files are available at http://www.mfix.org/members/download.php for Linux with ICC & PG and Windows with the following features:
      • All the new versions support inversion of colors (red - high and blue - low) by hitting the g-key.
      • The new ICC version supports files bigger than 2 GB.
      • The ICC and PG versions have a toggle switch ('toggle jpeg/tiff images') to toggle between jpeg and tiff output. This would be useful for any production quality snapshots with default set to JPG.
    • For the case of polydisperse particles (more than one solids phase), a new algorithm was used to compute ep_star if different particle diameters are used for the different solids phases. This new algorithm follows the work of Yu and Standish [Powder Tech., 52 (1987) 233-241] rather than the old implementation that followed the work of Fedors and Landel [Powder Tech., 23 (1979) 225-231]. The new algorithm is not limited to a binary mixture of powders (see calc_ep_star for details).
    • Several more updates for code clean up and for bug fixes - please refer to the CVS [5]


    New features

    • Direct Quadrature Method of Moments (DQMOM) - Contribution from Iowa State University [see 6 , 7, dqmom test case in mfix/tests directory and the readme file]. This change will need you to define D_p in mfix.dat as D_p0.
    • Time-splitting for chemistry - Contribution from Iowa State University [see 6, 8, isat test case in tests directory and the readme file].
    • In Situ Adaptive Tabulation (ISAT) - Contribution from Iowa State University [see 6, 8, isat test case in tests directory and the readme file].
    • Cohesion in Discrete Element Simulation (DES) - Contribution from University of Colorado [see 9, cohesion test cases in tests directory and the readme file].
    • Ahmadi and Simonin models for gas-solid turbulence - [see 10,  several tests and tutorial cases and the readme].


    References

    1. mfix/doc/”High resolution and conservation”
    2. mfix/doc/”Verification of momentum conservation”
    3. mfix/doc/” Normalization of correction equation residuals”
    4. mfix/doc/” Verification of the new residuals normalization method”
    5. http://www.mfix.org/cgi-bin/cvsweb.cgi/mfix/model/#dirlist
    6. mfix/doc/isat_dqmom
    7. http://www.mfix.org/open_citations/citation_page.php?citation_id=53
    8. http://www.mfix.org/open_citations/citation_page.php?citation_id=52
    9. http://www.mfix.org/open_citations/citation_page.php?citation_id=54
    10. mfix/doc/”Simonin_Ahmadi_Models.pdf”

    Back to Top

    March 4, 2005 "Plan for major structural changes in MFIX code"

    Dear Colleague:

    We have exciting news to share with you. We have been planning for some time to make major structural changes in MFIX code to make it a flexible and better research software. A summary of the anticipated major features are as follows:
    • Script-based front end to allow users to make major changes in the physical models and numerical techniques quickly and in an error-free manner
    • Component-based design to allow the inclusion of (or interoperability with) best-in-class open source software components developed at national labs and universities.
    • Open-source development process and infrastructure for software quality (accuracy and performance) assurance.
    We will continue to support the open-source model for disseminating information and for promoting research in computational multiphase flow.

    We have drafted a user requirements document (URD) <http://www.mfix.org/open_citations/uploads/1109885462URD%201.3.doc> and development effort has begun in March 2005. We will be periodically developing and posting task-lists which will provide much more specific information regarding our development plan.  The first task-list is expected by the end of June 2005.  Perhaps various tasks will align with your interests and could become a separate research project that can contribute to this development effort.  If you have any comments on this project or want any additional information, please email us at admin@lists.mfix.org.

    We will be sending out a survey form very soon to get input from you to determine what you would like to see in the next generation software. Your prompt response will contribute toward the success of this project.

    In the mean time, please be assured that we will keep supporting the current version of MFIX at the same-level as before but we will not be adding any new features beyond the upcoming release.

    Sincerely,

    MFIX team

    Back to Top

Disclaimers | Contact us | Feedback | Register
MFIX Home | About MFIX | Announcements | Partners | Documentations | Publications | Results | Links | Members Area | Open Citations
Send mail to admin@lists.mfix.org with questions, comments or suggestions about this web site.
Site created by Vitaly Prikhodko